Best Of GDC: Ultra Kill! Epic Talks Unreal Tournament III Development

Epic Games president Mike Capps began humbly by introducing the company as a developer of both games and middleware, even saying that “a lot of great stuff is developed on our middleware. Stuff like Bioshock, which puts Gears of War to shame.” Producer Jeff Morris introduced himself mentioning his time working at Firaxis, working on games including Civilization, but laughed that he “got tired of morally excusable games. I wanted to work on headshots and ultra kills.” They quickly moved on to discuss their intentions for Unreal Tournament 3: to keep the very fast (“zero time for spectacle”) gameplay, but to introduce the new game type, Warfare, add a new vehicle team, the Necris, improve the single player game immeasurably, with real characters (“not just some space clown that you’ve put together,” said Morris) offer modification possibilities even on consoles, and offer detailed character customization. What Went Right Knew the Genre: “This one, I think we did it exactly right,” said Capps. “And I think that was supported by our testing strategy. We’d test it constantly while building and revising levels.” Unified Look and Feel: “Visually I think it’s a real feast. You might have seen in the earlier Unreal titles, there wasn’t really a very consistent look. There’d be an afro chick fighting a robot in front of some crazy buildings,” admitted Capps. “There was something kind of cool about that, but I’m so proud of how, in particular, the Necris look like they fit the world.” Content Pipeline: Continued Capps, “We created 18 vehicles that were very different from anything we’d ever worked on before. They weren’t just jeeps and warthogs or whatever. Similarly with the weapons.” He added, “We spent a lot of time on cross-training, which was new for Epic --letting people swap around to different areas. The best results were finding the diamonds in the rough, like people who have been level design forever and would be absolutely great on something else. Some people would totally suck, though.” Mod Support: Capps noted that mods have been important to Epic from day one, with Tim Sweeney’s ZZT (“developed in his parents' basement”) and couldn’t help himself from hoping out loud that Microsoft would start to let mods be available on their closed system, and “be as cool about it as Sony was.” Unreal Engine 3: As a licensee of their own engine, Epic was able to receive all of the benefits of developing with the Unreal Engine that its other licensees gained, including easy integration of plugins and other software from their partners. What Went Wrong “I better preamble this by saying that Midway rocks,” began Morris, “so when I’m bashing the promotion of the game I’m not bashing them, I’m bashing the wrong route they went down.” User Interface: “We went with an interface that was driven by art,” Morris said, showing a short video of their “blue sky” wish, a slow (if pretty) interface with a large amount of animation. They next tried to speed it up, but still found themselves with an interface that was just too console oriented. In the end, out of time, they shipped with a very basic interface. “Fundamentally, the problem with the final UI was that we just didn’t have the time to polish it,” said Morris. Capps expanded, “We built in the time for polish with Unreal Tournament III, but the protracted development meant we ran out of time." Character Customization: “This was an extremely complicated thing to create and resulted in a very poor bang-for-buck. If you saw the APB demonstration yesterday, that’s a character customization system from the gods, but ours was so coarse most people couldn’t even tell what you’d done to customize your character!” Morris explained, “It had that fatal production flaw of starting asset creation before we started developing the system. I wish we could say we acknowledged the risk before we did that, but we were oblivious. The thing that absolutely blew the most was that it was not satisfying to use. Nobody even knew if you were a robot or a Necris in the first place.” Capps joked, “It was totally Jeff’s idea. I fought it the whole way.” Campaign: “We put a lot time on the cinematics, and we feel that those are some of the finest cinematics we’ve ever produced at Epic,” began Morris, before Capps interjected “except for the writing. The writing isn’t as good.” Making light of one of Unreal Tournament III’s main characters, Morris agreed with a gravelly “Word.” He described that the problem with the single player campaign, which they had very high hopes for, was that they just did not have the time to polish it. “It was good, but it wasn’t Bioshock” he said. In particular, they suffered from hyping the campaign heavily, and that “the story never clicked. People didn’t care about Reaper as a person, and a lot of the reason for that was that the length of time between cinematics was really long. It could go 4, 5 hours between them. Platform/Schedule: “Fundamentally,” Morris opined, “we ran out of polish time at the end. The problem was that we hyper-polished some aspects and didn’t polish others, and it was the case that the features we were comfortable with that we polished. All of the 1.0 features, character customization, the campaign -- that, we didn’t polish.” Scheduling offered just as many problems, with strung-out platform and region releases and a released that walked them into “the most FPS heavy Christmas of all time.” “It was a bloodbath,” lamented Capps. Promotion: “Midway got us every single [magazine] cover we went for,” admitted Morris, but what he felt was one of the biggest problems was that they “announced over two years before we shipped.” “You can’t talk about features year after year … We hyped story all 2007, and it was to its detriment. It’s better to announce everything six months before you ship, and people will find it hard to believe all you’ve included. We also talked a lot about features that were cut -- but which admitted deserved to be cut.”